Exhaustive bivariate Genome-Wide-Interaction-Studies applied to the uk-biobank datasets
Principal Investigator: Dr Clement Chatelain
Approved Research ID: 41906
Approval date: March 26th 2019
Today, thousands of genetic variants have been associated with diverse phenotypes, such as weight, cholesterol level, or risk of type 2 diabetes, giving insights into diseases mechanisms. However the impact in terms of new therapeutic approaches has been relatively limited. Moreover these variants independently explain only a small fraction of the estimated disease heritability. For example, in Crohn's Disease, the addition of the effects of associated variants explains 10.6% of the disease variability while the estimated heritability is 53%, and in Type-2 diabetes, 4.7% for an estimated heritability of 26%. Part of the missing heritability in Genome Wide Association Studies (GWAS) is expected to be explained by the fact that some phenotypes are not only driven by genetic variants acting independently but also acting in interaction, also called epistasis. We have implemented and compared the most popular algorithms that perform genome-wide scans for epistasis in GWAS. In a recent work we have estimated that GWAS with more than 10,000 individuals are required to detect moderate epistasis with sufficient power. With over half million individual with genetic data and extensive phenotypic information the UK biobank is therefore an ideal tool to better understand how epistasis shape the relationship between our genotype and phenotype. The objective of this project is to perform genome-wide epistasis scan for more than 1000 selected phenotypes using high performance computing facilities. The results of all these analysis will be shared with the UK biobank and the scientific community. We estimate the duration of the full project to be about 12 months. As a large pharmaceutical company Sanofi main objective is the marketing of effective and safe medicines for human health. In the process of developing new drugs, the final validation of the therapeutic interest of a therapeutic target is made only from the clinical phase 2 and 3. It is therefore important to choose the target well at the beginning of the research stage, to limit the attrition during these 2 clinical phases and speed up the availability of new medicines for patients. The use of human genetic data is one of the approaches we use to increase the chances of success of our programs. Moreover, by sharing all the results with the scientific community we also give the opportunity to other companies or academic labs to exploit those new data for their own scientific research.